Multi-dimensional Bayesian network classifiers: A survey

Gil-Begue, Santiago; Bielza, Concha; Larrañaga, Pedro

doi:10.1007/s10462-020-09858-x

Cited by 27 publications

(24 citation statements)

References 98 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Traditional equations for precision, recall and, therefore,

F_{1}

score can only be used for a unique binary class variable. However, Gil‐Begue et al 53 extended them for multiple, possibly nonbinary, class variables. Let

B

be a function that computes any of these evaluation metrics by receiving a confusion matrix, then the metric scores are obtained with macro‐ and micro‐averaging as follows: Macro‐averaging: averages the scores of each class variable:

B_{m a c r o} = \frac{1}{d} \sum_{y}^{d} B_{C_{y}}, where B_{C_{y}} = ({, \begin{matrix} \frac{1}{∣ Ω_{C_{y}} ∣} \sum_{c_{j}} B ({t p}_{c_{j}}, {f p}_{c_{j}}, {t n}_{c_{j}}, {f n}_{c_{j}}), & if ∣ Ω_{C_{y}} ∣ > 2 \\ B ({t p}_{C_{y}}, {f p}_{C_{y}}, {t n}_{C_{y}}, {f n}_{C_{y}}) & otherwise . \end{matrix})

If the class variable

C_{y}

is binary, only the confusion matrix for one of its classes (

{t p}_{C_{y}}, {f p}_{C_{y}}, {t n}_{C_{y}}, {f n}_{C_{y}}

) is considered. Micro‐averaging: aggregates the confusion matrices of each class variable:

B_{m i c r o} = B ()(, \sum_{y = 1}^{d} ...)

…”

Section: Methodsmentioning

confidence: 99%

“…Traditional equations for precision, recall and, therefore, F 1 score can only be used for a unique binary class variable. However, Gil-Begue et al 53 extended them for multiple, possibly nonbinary, class variables. Let B be a function that computes any of these evaluation metrics by receiving a confusion matrix, then the metric scores are obtained with macro-and microaveraging as follows:…”

mentioning

confidence: 99%

See 1 more Smart Citation

Multidimensional continuous time Bayesian network classifiers

Villa-Blanco

Larrañaga

Bielza

2021

Int J of Intelligent Sys

Self Cite

View full text Add to dashboard Cite

The multidimensional classification of multivariate time series deals with the assignment of multiple classes to time-ordered data described by a set of feature variables. Although this challenging task has received almost no attention in the literature, it is present in a wide variety of domains, such as medicine, finance or industry. The complexity of this problem lies in two nontrivial tasks, the learning with multivariate time series in continuous time and the simultaneous classification of multiple class variables that may show dependencies between them. These can be addressed with different strategies, but most of them may involve a difficult preprocessing of the data, high space and classification complexity or ignoring useful interclass dependencies. Additionally, no attention has been given to the development of new multidimensional classifiers of time series based on probabilistic graphical models, even though transparent models can facilitate further understanding of the domain. In this paper, a novel probabilistic graphical model is proposed, which is able to classify a discrete multivariate temporal sequence into multiple class variables while modeling their dependencies. This model extends continuous time Bayesian networks to the multidimensional classification problem, which are able to explicitly represent the behavior of time series that This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

“…Traditional equations for precision, recall and, therefore,

F_{1}

score can only be used for a unique binary class variable. However, Gil‐Begue et al 53 extended them for multiple, possibly nonbinary, class variables. Let

B

B_{m a c r o} = \frac{1}{d} \sum_{y}^{d} B_{C_{y}}, where B_{C_{y}} = ({, \begin{matrix} \frac{1}{∣ Ω_{C_{y}} ∣} \sum_{c_{j}} B ({t p}_{c_{j}}, {f p}_{c_{j}}, {t n}_{c_{j}}, {f n}_{c_{j}}), & if ∣ Ω_{C_{y}} ∣ > 2 \\ B ({t p}_{C_{y}}, {f p}_{C_{y}}, {t n}_{C_{y}}, {f n}_{C_{y}}) & otherwise . \end{matrix})

If the class variable

C_{y}

is binary, only the confusion matrix for one of its classes (

{t p}_{C_{y}}, {f p}_{C_{y}}, {t n}_{C_{y}}, {f n}_{C_{y}}

) is considered. Micro‐averaging: aggregates the confusion matrices of each class variable:

B_{m i c r o} = B ()(, \sum_{y = 1}^{d} ...)

…”

Section: Methodsmentioning

confidence: 99%

mentioning

confidence: 99%

Multidimensional continuous time Bayesian network classifiers

Villa-Blanco

Larrañaga

Bielza

2021

Int J of Intelligent Sys

Self Cite

View full text Add to dashboard Cite

show abstract

“…Bayes Net was used as a classifier and can lead to extremely precise categories if correctly educated. A comprehensive study can be discovered in [ 51 , 52 , 53 ] on the Bayesian Network.…”

Section: Background and Related Workmentioning

confidence: 99%

Predicting Attack Pattern via Machine Learning by Exploiting Stateful Firewall as Virtual Network Function in an SDN Network

Prabakaran

Ramalakshmi

Hussain

et al. 2022

Sensors

View full text Add to dashboard Cite

Decoupled data and control planes in Software Defined Networks (SDN) allow them to handle an increasing number of threats by limiting harmful network links at the switching stage. As storage, high-end servers, and network devices, Network Function Virtualization (NFV) is designed to replace purpose-built network elements with VNFs (Virtualized Network Functions). A Software Defined Network Function Virtualization (SDNFV) network is designed in this paper to boost network performance. Stateful firewall services are deployed as VNFs in the SDN network in this article to offer security and boost network scalability. The SDN controller’s role is to develop a set of guidelines and rules to avoid hazardous network connectivity. Intruder assaults that employ numerous socket addresses cannot be adequately protected by these strategies. Machine learning algorithms are trained using traditional network threat intelligence data to identify potentially malicious linkages and probable attack targets. Based on conventional network data (DT), Bayesian Network (BayesNet), Naive-Bayes, C4.5, and Decision Table (DT) algorithms are used to predict the target host that will be attacked. The experimental results shows that the Bayesian Network algorithm achieved an average prediction accuracy of 92.87%, Native–Bayes Algorithm achieved an average prediction accuracy of 87.81%, C4.5 Algorithm achieved an average prediction accuracy of 84.92%, and the Decision Tree algorithm achieved an average prediction accuracy of 83.18%. There were 451 k login attempts from 178 different countries, with over 70 k source IP addresses and 40 k source port addresses recorded in a large dataset from nine honeypot servers.

show abstract

“…However, the combinatorial nature still exists, which leads to that the deficiencies cannot be fully addressed. A family of MDC models called multi-dimensional Bayesian network classifier [25] aims at learning different kinds of DAG structures over class spaces to explicitly model the class dependencies. However, determining DAG structures is computationally demanding, and only nominal features can be tackled generally.…”

Section: Related Workmentioning

confidence: 99%

Multi-dimensional Classification via Selective Feature Augmentation

Jia

Zhang

2022

Mach. Intell. Res.

View full text Add to dashboard Cite

In multi-dimensional classification (MDC), the semantics of objects are characterized by multiple class spaces from different dimensions. Most MDC approaches try to explicitly model the dependencies among class spaces in output space. In contrast, the recently proposed feature augmentation strategy, which aims at manipulating feature space, has also been shown to be an effective solution for MDC. However, existing feature augmentation approaches only focus on designing holistic augmented features to be appended with the original features, while better generalization performance could be achieved by exploiting multiple kinds of augmented features. In this paper, we propose the selective feature augmentation strategy that focuses on synergizing multiple kinds of augmented features. Specifically, by assuming that only part of the augmented features is pertinent and useful for each dimension’s model induction, we derive a classification model which can fully utilize the original features while conduct feature selection for the augmented features. To validate the effectiveness of the proposed strategy, we generate three kinds of simple augmented features based on standard kNN, weighted kNN, and maximum margin techniques, respectively. Comparative studies show that the proposed strategy achieves superior performance against both state-of-the-art MDC approaches and its degenerated versions with either kind of augmented features.

show abstract

Multi-dimensional Bayesian network classifiers: A survey

Cited by 27 publications

References 98 publications

Multidimensional continuous time Bayesian network classifiers

Multidimensional continuous time Bayesian network classifiers

Predicting Attack Pattern via Machine Learning by Exploiting Stateful Firewall as Virtual Network Function in an SDN Network

Multi-dimensional Classification via Selective Feature Augmentation

Contact Info

Product

Resources

About